Create first working implementation

[TiSiE]

I've created a gist to demonstrate a sample of how the classification could work.

This API needs at least one entry points to classify input strings, and another to train / seed data (via uploading CSV-files.)

There are different attributes for a job ad to be classified, such as employment type or language

The API entry points should be

• /classify/[attribute]
  POST-method : Will consume a json payload and return a json result.

  {
  "title": "the job title (required)",
  "tasks": "job tasks text",
  "qualifications": "job qualifications text",
  "benefits": "job benefits text"
  }

  {
  "label": "classified label"
  }

• /train/[attribute] POST-method: Will consume a CSV data in the post request body, store it, use it to train the classifier. returns a json result.

  {
  status: "success|failed",
  message: "error message (optional)"
  }

---

This is a first draft and is open to suggestions.

The entry points urls could also be

• [attribute]/
• [attribute]/train

There's no need to use JSON, if there are other methods better fitted for this use that I'm not aware of.

Leave your thoughts in the comments.

[cbleek]

Maybe we should use https://github.com/RubixML/RubixML for the first implementation